Handling Delayed Feedback in Distributed Online Optimization: A Projection-Free Approach - Systèmes Répartis, Calcul Parallèle et Réseaux
Communication Dans Un Congrès Année : 2024

Handling Delayed Feedback in Distributed Online Optimization: A Projection-Free Approach

Résumé

Learning at the edges has become increasingly important as large quantities of data are continually generated locally. Among others, this paradigm requires algorithms that are \emph{simple} (so that they can be executed by local devices), \emph{robust} (again uncertainty as data are continually generated), and \emph{reliable} in a distributed manner under network issues, especially delays. In this study, we investigate the problem of online convex optimization (\oco) under adversarial delayed feedback. We propose two projection-free algorithms for centralized and distributed settings in which they are carefully designed to achieve a regret bound of $O(\sqrt{B})$ where $B$ is the sum of delay, which is optimal for the OCO problem in the delay setting while still being projection-free. We provide an extensive theoretical study and experimentally validate the performance of our algorithms by comparing them with existing ones on real-world problems.
Fichier principal
Vignette du fichier
2402.02114v1.pdf (806.94 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04703453 , version 1 (20-09-2024)

Licence

Identifiants

Citer

Tuan-Anh Nguyen, Nguyen Kim Thang, Denis Trystram. Handling Delayed Feedback in Distributed Online Optimization: A Projection-Free Approach. Joint European Conference on Machine Learning and Knowledge Discovery in Databases - ECML PKDD 2024, Sep 2024, Vilnius, Lithuania. pp.197-211, ⟨10.1007/978-3-031-70341-6_12⟩. ⟨hal-04703453⟩
0 Consultations
0 Téléchargements

Altmetric

Partager

More